Alzheimer's disease (AD) and non-AD dementias are increasing in frequency, cause major disability, and remain largely untreatable. They may be caused by the accumulation of distinct proteins with pathogenic conformations. Transgenic mice expressing human amyloid precursor proteins (hAPP) accumulate amyloid-beta peptides (Abeta) in their brain and develop pathological and behavioral alterations resembling AD. However, the mechanisms underlying cognitive deficits in AD and in hAPP mice are largely unknown. There also is an urgent need, from both a diagnostic and therapeutic perspective, for the identification of reliable biochemical markers of Abeta-induced cognitive decline. We have pinpointed molecular alterations that accurately reflect and possibly contribute to AD-related cognitive impairments. Deficits in spatial learning and memory in mice expressing familial AD-mutant hAPP correlated most robustly with reductions in the calcium-binding protein calbindin-D28k in granule cells of the dentate gyrus, a brain region critically involved in learning and memory. These reductions were age-dependent and correlated with the relative abundance of Abeta1-42 but not with the amount of Abeta deposited in amyloid plaques. In our preliminary studies, we have identified strong reductions in calbindin levels also in granule cells of humans with AD with the greatest reductions seen in the most severely demented cases. In Specific Aim 1, we will determine whether calbindin reductions in granule cells of the dentate gyrus correlate with specific cognitive deficits (e.g., in spatial learning and memory) and genetic factors (e.g., APOE haplotype) in humans with AD and non-AD dementias. In Specific Aim 2, we will examine in these cases how calbindin reductions in granule cells relate to the presence and severity of other neuropathological alterations and to the accumulation of Abeta and other neurotoxic proteins in different brain regions. In Specific Aim 3, we will analyze hAPP transgenic mice with oligonucleotide arrays and protein arrays to search for additional molecular indicators and mediators of Abeta-dependent behavioral deficits. The clinical relevance and usefulness of the most interesting molecular alterations will then be evaluated in human subjects, either in the context of this ADRC or in independent studies. We predict that this Project will shed light on the pathways leading from the accumulation of Abeta to cognitive decline, ascertain the usefulness of a promising indicator of Abeta-induced neuronal deficits, and pinpoint novel markers and mediators of pathological cognitive decline in aging. The success of this project depends critically on the support of the proposed cores and on the overall environment this ADRC would create at UCSF.